JP2016065750A - Underwater positioning system and underwater positioning method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve accuracy of underwater positioning without grasping a numerical value of an error factor.SOLUTION: An underwater positioning system calculates at least any one of a coupling load and a threshold in a neural network, by comparing a position that is output from an output layer in a case where for an underwater known point, a position measured by positioning means is input to an input layer in the neural network, and a position of the known position; and provides, as a positioning result, a position that is output from the output layer in a case where for an underwater unknown point, a position measured by positioning means is input to the input layer in the neural network.SELECTED DRAWING: Figure 3

Description

  The present invention relates to a technique for measuring the position of an underwater structure.

  When submerging an underwater structure such as a caisson, it is necessary to accurately measure the position of the underwater structure during the installation operation. The position of the underwater structure is measured by the following configuration, for example. A transponder is installed in the underwater structure, and an acoustic positioning device and a satellite positioning device are installed in the hoist ship that suspends the underwater structure. The acoustic positioning device includes a transmitter and a receiver array, and calculates the distance from the transmission of the ultrasonic signal to the reception and the distance from the underwater sound speed to the transponder to form the receiver array. The direction of the transponder is calculated from the phase difference of the signals received by each receiver. This distance and direction indicate the relative position of the transponder with respect to the acoustic positioning device. The satellite positioning device calculates the absolute position of its own device using the signal received from the navigation satellite. The absolute position of the transponder is calculated by converting the relative position of the transponder into an absolute position using this absolute position as a reference position. Patent Document 1 discloses a technique for sinking an underwater structure using this type of positioning technique.

JP 2010-229656 A

By the way, in positioning in water using ultrasonic waves, errors may occur due to various causes such as water temperature, water pressure, underwater electrical conductivity, underwater salinity, underwater sound speed, and the like. In order to improve the accuracy of positioning, it is desirable to correct the error based on the numerical values of these causes, but in reality, these numerical values may vary depending on the position in the water, and change over time. In some cases, it is difficult to measure these numerical values in detail.
Therefore, the present invention provides a technique that can improve the accuracy of underwater positioning without knowing the numerical value of the cause of the error.

  The present invention provides a positioning means for measuring a position in water, a position output from an output layer when the position measured by the positioning means for a known point in water is input to an input layer in a neural network, A calculating means for calculating at least one of a coupling load or a threshold value in the neural network by comparing the position and a position measured by the positioning means for an unknown point in water is input to an input layer in the neural network An underwater positioning system is provided that includes providing means for providing the position output from the output layer as a positioning result in the case of the above.

In the above configuration, the calculating means includes at least one of water temperature, water pressure, underwater electrical conductivity, underwater salinity and underwater sound speed in addition to the position measured by the positioning means with respect to a known point in water. By comparing the position output from the output layer when input to the input layer in the neural network and the position of the known point, at least one of the coupling load and the threshold value may be calculated.
In the above configuration, the calculation means may calculate at least one of the combined load and the threshold every time positioning is repeated a plurality of times for a known point in water.
In the above configuration, when the conditions regarding the location and time of the previous and current positioning are included in a predetermined range, the calculating means calculates the combined load or the threshold value calculated based on the previous positioning as the current time. It may be used as the initial value of the combined load or the threshold value at the time of positioning.

  Further, the present invention compares the position output from the output layer with the position of the known point when the position measured by the positioning means for the known point in water is input to the input layer in the neural network, A calculation step for calculating at least one of a coupling load and a threshold value in the neural network and a position measured by the positioning means for an unknown point in water are output from the output layer when the position is input to the input layer in the neural network. An underwater positioning method comprising: a providing step of providing a position as a positioning result.

  According to the present invention, the accuracy of underwater positioning can be improved without grasping the numerical value of the cause of the error.

1 is a diagram illustrating an overall configuration of an underwater positioning system 1. FIG. It is a figure which shows the hardware constitutions of the underwater positioning system. It is a flowchart which shows the process which the computer 5 performs. It is a figure which shows the hierarchical structure model of a neural network. It is a figure showing underwater positioning system 1A.

  An example for carrying out the present invention will be described. In the present embodiment, it is assumed that an underwater structure (hereinafter referred to as a new structure 200) is newly installed in the vicinity of an existing underwater structure (hereinafter referred to as an existing structure 100). The underwater structure may be any caisson, pile, block, or the like.

  FIG. 1 is a diagram showing the overall configuration of the underwater positioning system 1. A new structure 200 is suspended from a boom 92 of a hoist 91 provided in the hoist ship 9 by a wire rope 93. The underwater positioning system 1 includes an acoustic positioning device 2, a satellite positioning device 3, a CTD (Conductivity Temperature Depth profiler) 4, a computer 5, an existing side transponder 71, and a new side transponder 72. The acoustic positioning device 2, the satellite positioning device 3, and the CTD 4 are connected to the computer 5 by communication means.

  The acoustic positioning device 2, the satellite positioning device 3, the CTD 4, and the computer 5 are provided on the hoist ship 9. In the following description, the existing side transponder 71 and the new side transponder 72 are collectively referred to as a transponder 7. The existing structure 100 and the new structure 200 are provided with a pedestal for attaching / detaching the transponder 7 in advance (not shown).

  The existing structure 100 is provided with two existing transponders 71. For example, as shown in FIG. 1, one existing side transponder 71 is provided at each of two locations at the top end of the existing structure 100. The new structure 200 is provided with three new-side transponders 72. For example, as shown in FIG. 1, three new-side transponders 72 are provided so that a triangle whose apex is the installation position of the new-side transponder 72 is formed at the top end of the new structure 200. By measuring the positions of the three new-side transponders 72, the position, orientation, and inclination of the new structure 200 are obtained.

  The acoustic positioning device 2 is an example of a positioning unit that measures a position in water. The acoustic positioning device 2 is, for example, an SSBL (Super Short Base Line) system device including a single transmitter and a receiver array including a plurality of receivers. The transmitter transmits an ultrasonic interrogation signal, and the transponder 7 receiving the interrogation signal transmits an ultrasonic response signal. When the receiver array receives this response signal, the acoustic positioning device 2 calculates the distance from the transmission to the reception and the distance from the underwater sound speed to the transponder 7 and receives each reception of the receiver array. The direction of the transponder 7 is calculated from the phase difference of the signal received by the waver. The relative position of the transponder 7 with respect to the acoustic positioning device 2 is indicated by this distance and direction. The acoustic positioning device 2 transmits the calculated relative position of the transponder 7 to the computer 5.

The satellite positioning device 3 calculates the absolute position of the own device using a signal received from a navigation satellite of GNSS (Global Navigation Satellite System), and transmits the calculated absolute position to the computer 5.
The CTD 4 measures underwater electrical conductivity, water temperature, and water depth (water pressure), and transmits these measured values to the computer 5.

  The computer 5 is installed in the operation room of the hoist 91. The computer 5 converts the relative position of the transponder 7 received from the acoustic positioning device 2 into an absolute position by using the absolute position of the satellite positioning device 3 received from the satellite positioning device 3 as a reference position, thereby obtaining the absolute position of the transponder 7. calculate. In addition, the computer 5 executes processing by a neural network (ANN: Artificial Neural Network) using the calculated absolute position of the transponder 7 and the measured value received from the CTD 4 as input values. Details of the processing by the neural network will be described later.

  FIG. 2 is a diagram illustrating a hardware configuration of the underwater positioning system 1. The computer 5 includes a control unit 51, a storage unit 52, and an external IF unit 53. The control unit 51 includes an arithmetic device such as a CPU (Central Processing Unit) and a storage device such as a ROM (Read Only Memory) and a RAM (Random Access Memory) (all not shown). The ROM stores firmware that describes the procedure for starting up hardware and an OS (Operating System). The RAM is used for storing data when the CPU executes an operation.

  The storage unit 52 includes, for example, a hard disk storage device, and stores an OS, application programs, and the like. The external IF (Interface) unit 53 is connected to the acoustic positioning device 2, the satellite positioning device 3, the CTD 4, the input device 54, the display device 55, and the printing device 56. The input device 54 is, for example, a keyboard or a mouse. The display device 55 is, for example, a liquid crystal display panel. The printing device 56 is, for example, an ink jet printer.

  FIG. 3 is a flowchart showing processing executed by the computer 5. The control unit 51 of the computer 5 executes this process according to the procedure described by the application program stored in the storage unit 52. In this processing, the known point is the existing transponder 71 provided in the existing structure 100, and its absolute position is known. The unknown point is the new-side transponder 72 provided in the new structure 200, and its absolute position is unknown.

  When the position of an underwater structure is measured by the acoustic positioning device 2, an error occurs in the measured value due to various causes such as water temperature, water pressure, underwater electrical conductivity, underwater salinity, underwater sound velocity, etc. May not match the true absolute position. In the present embodiment, in order to correct this error, the absolute position obtained by measuring the known point is input to the neural network, and the absolute position output from the neural network is input using the known absolute position of the known point as a teacher signal. And the teacher signal are compared, and the coupling load of the neural network is corrected so that the absolute position output from the neural network approaches the teacher signal. This process is called learning. Then, the absolute position obtained by measuring the unknown point is input to the neural network having the coupling weight corrected by learning, and the output value is provided as the positioning result.

  FIG. 4 is a diagram showing a hierarchical structure model of a neural network. The neural network in the present embodiment is a three-layer hierarchical model including an input layer, an intermediate layer, and an output layer, and uses a sigmoid function as a response function.

  In step S11, the control unit 51 inputs a measured value of a known point to the neural network. Specifically, the control unit 51 inputs the absolute position of the existing transponder 71 and the measurement value received from the CTD 4 to the input layer. This absolute position is an absolute position calculated from the relative position of the existing transponder 71 received from the acoustic positioning device 2 and the absolute position of the satellite positioning device 3 received from the satellite positioning device 3. In the present embodiment, the absolute position of the existing transponder 71 is represented by coordinate values X, Y, and Z in the three-dimensional orthogonal coordinate system, and input to the input layer as values I1, I2, and I3, respectively. FIG. 4 shows an example in which one type of measured values (underwater electrical conductivity, water temperature, water depth) received from the CTD 4 is input to the input layer as a value I4. , I5, or three types of measurement values may be input as values I4, I5, and I6.

In step S15, the control unit 51 obtains an intermediate layer error from the output layer error and the intermediate layer value, and corrects the coupling load between the input layer and the intermediate layer. Similar to step S14, the differential of the error evaluation scale is expressed by the following equation.
The control unit 51 corrects the combined load by the following equation.
That is, the control unit 51 compares the position output from the output layer with the position of the known point when the position measured by the positioning unit for the known point in water is input to the input layer in the neural network, It has a function as a calculation means for calculating at least one of the coupling load and the threshold value in the neural network. In addition, the control unit 51 calculates at least one of the combined load and the threshold every time the positioning is repeated a plurality of times for a known point in water.

  In step S16, when the error evaluation scale falls below the threshold value, the control unit 51 determines that learning may be ended (step S16: YES), and proceeds to the process of step S21. On the other hand, if the error evaluation scale is not below the threshold value, it is determined that iterative learning is necessary (step S16: NO), and the process returns to step S11.

  In step S21, the control unit 51 inputs the measurement value of the unknown point to the neural network. Specifically, the control unit 51 inputs the absolute position of the newly installed transponder and the measurement value received from the CTD to the input layer. This absolute position is an absolute position calculated from the relative position of the new transponder 72 received from the acoustic positioning device 2 and the absolute position of the satellite positioning device 3 received from the satellite positioning device 3.

In step S22, the control unit 51 outputs the absolute position of the unknown point obtained by the neural network. Specifically, the control unit 51 causes the display device 55 to display an image indicating the absolute position obtained by the neural network. That is, the control unit 51 has a function as a providing unit that provides a position output from the output layer as a positioning result when a position measured by the positioning unit for an unknown point in water is input to the input layer in the neural network.
A worker who operates the hoist 91 confirms the absolute position of the unknown point by viewing this image, and operates the hoist 91 so that the new structure 200 is brought closer to the target position. When the installation of the new structure 200 is completed, the worker inputs an instruction to end the positioning to the computer 5.

  In step S23, the control unit 51 determines whether or not to end positioning. Specifically, when an instruction to end the positioning is input (step S22: YES), the control unit 51 ends the process and does not input an instruction to end the positioning ( In step S22: NO), the process returns to step S21. That is, until the instruction to end the positioning is input to the computer, the control unit 51 repeats the processing from step S21 to S23, so that the absolute position of the new structure 200 is transmitted to the worker in real time. .

<Modification>
The above embodiment may be modified as follows. A plurality of modified examples may be combined.
<1>
The acoustic positioning device 2 may be provided in addition to the hoist ship 9. For example, FIG. 5 is a diagram illustrating an underwater positioning system 1 </ b> A in which the acoustic positioning device 2 is provided in the new structure 200. In this case, the tower 21 is installed at the top of the new structure 200 and the acoustic positioning device 2 is installed at the upper end of the tower 21 so that transmission and reception with the existing transponder 71 are not blocked. . The absolute position of the acoustic positioning device 2 is measured by another positioning means. For example, a tower whose height is exposed at the top of the water surface may be installed at the top of the new structure 200, and the absolute position may be obtained by the satellite positioning device 3 installed at the top of the tower.

<2>
The CTD 4 may be provided on any of the existing structure 100, the new structure 200, and the seabed.
CTD4 may be omitted. In this case, the measured value of CTD4 is not input to the neural network.
One existing transponder 71 may be provided, or three or more existing transponders 71 may be provided. Four or more new-side transponders 72 may be provided.

<3>
The computer 5 may calculate an underwater salinity from the underwater electrical conductivity, the water temperature, and the water depth (water pressure) measured by the CTD 4, and input the underwater salinity to the neural network. Alternatively, the underwater sound speed may be measured using a sound speed / pressure sensor (SVPS) or the like instead of the CTD 4 and the underwater sound speed may be input to the neural network. Moreover, all of underwater electrical conductivity, water temperature, water depth (water pressure), underwater salinity, and underwater sound velocity may be input to the neural network, or at least one of these may be input to the neural network. . In short, in addition to the position where the positioning means has measured by ultrasonic waves at a known point in water, at least one of water temperature, water pressure, underwater electrical conductivity, underwater salinity and underwater sound speed is input to the input layer in the neural network. When the input is input to the output layer, the position output from the output layer and the position of the known point may be compared to calculate at least one of the coupling load and the threshold value.

<4>
The acoustic positioning method may be SBL (Short Base Line), USBL (Ultra Short Base Line), LBL (Long Base Line), or the like.

<5>
In general, in a neural network, sufficient learning is performed in advance, and positioning is performed based on the result. However, since the underwater positioning system 1 according to the present invention is premised on operation for construction, the environmental conditions are the same as when learning. It may be different from the time of positioning. Therefore, learning may be repeated until positioning is completed. For example, when the determination result of step S23 in the flowchart of FIG. 3 is NO, the process may return to step S11. Or you may make it perform the process from step S11 to step S16 regularly. Alternatively, when an operator operating the hoist 91 instructs the computer 5 to execute learning, the processing from step S11 to step S16 may be executed.

<6>
If conditions related to the location and time of the previous and current positioning are within the specified range, the combined load or threshold calculated based on the previous positioning is used as the initial value of the combined load or threshold at the current positioning. It may be used. For example, when the location and time of the previous and current positioning are close, it is conceivable that the numerical values such as water temperature, water pressure, underwater electrical conductivity, and underwater salinity are similar. In such a case, it is expected that the correction of the combined load or threshold value is completed faster when the previous combined load or threshold value is used as the initial value than when the combined load or threshold value is initialized in the current positioning.

<7>
The number of layers of the hierarchical structure model of the neural network may be any number. That is, a hierarchical structure model provided with a plurality of intermediate layers may be used.

1 Underwater positioning system, 2 Acoustic positioning device, 3 Satellite positioning device, 4 CTD, 5 Computer, 7 Transponder, 71 Existing transponder, 72 New transponder, 9 Hoist ship, 91 Hoist, 92 Boom, 93 Wire rope, 100 Existing Structure, 200 New structure

Claims (5)

Positioning means for measuring the position in water;
When the position measured by the positioning means for a known point in the water is input to the input layer in the neural network, the position output from the output layer is compared with the position of the known point to thereby determine the combined load in the neural network. Or calculating means for calculating at least one of the threshold values;
An underwater positioning system comprising: a providing means for providing a position output from an output layer as a positioning result when a position measured by the positioning means for an unknown point in water is input to an input layer in the neural network. The calculation means includes an input layer in the neural network that includes at least one of water temperature, water pressure, underwater electrical conductivity, underwater salinity, and underwater sound velocity in addition to the position measured by the positioning means with respect to a known point in water. 2. The underwater positioning system according to claim 1, wherein at least one of the combined load and the threshold value is calculated by comparing a position output from an output layer when input to the position and a position of the known point. . The underwater positioning system according to claim 1, wherein the calculating unit calculates at least one of the combined load and the threshold every time positioning is repeated a plurality of times for a known point in water. When the conditions regarding the location and time of the previous and current positioning are included within a predetermined range, the calculating means calculates the combined load or the threshold value calculated based on the previous positioning at the time of the current positioning. The underwater positioning system according to claim 1, wherein the underwater positioning system is used as a combined load or an initial value of the threshold value. By comparing the position output from the output layer when the position measured by the positioning means for the known point in water is input to the input layer in the neural network and the position of the known point, A calculating step for calculating at least one of the threshold values;
An underwater positioning method comprising: providing a position output from an output layer as a positioning result when a position measured by the positioning means for an unknown point in water is input to an input layer in the neural network.